About Projects 关于项目
==============
在大多数情况下，一个项目就是为一个网站编写的脚本
In most cases, a project is one script you write for one website.项目是独立的，但是您可以将另一个项目用`from projects import other_project`的方式导入为模块
* Projects are independent, but you can import another project as a module with `from projects import other_project`
一个项目有五种状态：`TODO`, `STOP`, `CHECKING`, `DEBUG` and `RUNNING`
* A project has 5 status: `TODO`, `STOP`, `CHECKING`, `DEBUG` and `RUNNING`TODO  刚刚创建一个可编辑的脚本- `TODO` - a script is just created to be writtenSTOP  如果您希望项目停止，您可以将项目标记为'STOP'- `STOP` - you can mark a project as `STOP` if you want it to STOP (= =).CHECKING 当正在运行的项目被修改时，为了防止不完整的修改，项目状态将被自动设置为“CHECKING”。- `CHECKING` - when a running project is modified, to prevent incomplete modification, project status will be set as `CHECKING` automatically.DEBUG`/`RUNNING 这两种状态对spider没有区别。但是当它第一次运行时，将它标记为‘DEBUG’，然后在检查之后将其更改为‘RUNNING’。- `DEBUG`/`RUNNING` - these two status have no difference to spider. But it's good to mark it as `DEBUG` when it's running the first time then change it to `RUNNING` after being checked.
爬行速度控制是通过`rate`和`burst`与令牌桶]与[token-bucket]（http://en.wikipedia.org/wiki/token_bucket）算法
* The crawl rate is controlled by `rate` and `burst` with [token-bucket](http://en.wikipedia.org/wiki/Token_bucket) algorithm.rate 表示一秒钟有多少请求- `rate` - how many requests in one secondburst 考虑到这种情况,`rate/burst = 0.1/3`,意思是每10s爬取1个网页。当所有任务都完成后，项目每分钟都在检查最后的更新列表。假设有3个新列表,pyspider将“burst”爬3个任务无需等待3×10秒，然而，第四任务需要等待10秒- `burst` - consider this situation, `rate/burst = 0.1/3`, it means that the spider scrawls 1 page every 10 seconds. All tasks are finished, project is checking last updated items every minute. Assume that 3 new items are found, pyspider will "burst" and crawl 3 tasks without waiting 3*10 seconds. However, the fourth task needs wait 10 seconds.
若要删除一个项目，请将“group”设置为“delete”并将状态设置为“STOP”，并等待24小时。
* To delete a project, set `group` to `delete` and status to `STOP`, wait 24 hours.`on_finished` callback
--------------------
你可以在项目中重写“on_finished”方法，当task_queue变成0时该方法将被触发
You can override `on_finished` method in the project, the method would be triggered when the task_queue goes to 0.
例子1：当你开始一个项目去爬取一个网站的100个页面，当100个页面爬取成功或者重试失败的时候，"on_finished"回调方法将被执行
Example 1: When you start a project to crawl a website with 100 pages, the `on_finished` callback will be fired when 100 pages are successfully crawled or failed after retries.
例子2：一个项目在“auto_recrawl”任务时，“on_finished”回调将不会触发，因为当auto_recrawl任务存在时，时间队列不可能变为0.
Example 2: A project with `auto_recrawl` tasks will **NEVER** trigger the `on_finished` callback, because time queue will never become 0 when there are auto_recrawl tasks in it.
例子3：一个项目在在每次新任务提交的时候通过“@every”装饰方法将触发“on_finished”回调。
Example 3: A project with `@every` decorated method will trigger the `on_finished` callback every time when the newly submitted tasks are finished.